Quantum Engineering Colloquium


1. Gautham Rangasamy,  TU Delft, The Netherlands -- 08-02-2017

2. Johan Peltenburg, TU Delft, The Netherlands -- 08-02-2017


1.    An Introduction to Monolithic Integration of Qubits and CMOS

1. Why do we need monolithic integration?
2. What approaches are currently being pursued in this direction?
3. Results so far
4. Future plans and Conclusions 

 2.    Towards a User-Transparent Approach to Integrate Low-Level Accelerators with a High Level Cluster-Computing Framework

As the amount of data that people and their devices produce increases at a staggering rate, the field of Big Data has seen a lot of interest over the past few years. Many scientists, companies and governments are very interested in processing this data for different purposes. Nowadays, processing this data is still often done on clusters of commodity hardware; the nodes contain general-purpose machines, which often use a multi-core approach to process the data. It is predicted that future performance within this paradigm is limited (Dark Silicon), while the rate at which the data grows only increases. This calls for the need to integrate other paradigms (such as GPGPU or FPGA accelerated computing) with the frameworks being used by big data analysts.
In this talk, we will take a look at one of such frameworks (Spark) that allows for fault-tolerant cluster computing on commodity hardware. Furthermore, we will discuss some challenges and opportunities to integrate accelerators with this framework, while trying to keep the big data analysts oblivious to the side-issues related to the low-level implementation.



CE Tweets